Is finance responsible for the rise in wage inequalities in France

Is finance responsible for the rise in wage inequality
in France?
Abstract
Based on the DADS, a very detailed French database on wages, we show that
wage inequality started to increase in France in the mid-1990s. This phenomenon is limited to the top end of income distribution and concerns mainly the
top 0.1%, whose share of total salaries increased from 1.2% to 2% between
1996 and 2007. This increase in inequality was accompanied by some changes
in the social composition of this wage elite. These include a decline in CEOs
and an increase in lower rank management, such as chief officers and other
administrative managers, as well as a rise in sportspersons. A sector approach
shows that finance (3% of private sector employees) is responsible for half of
the rise in inequality at the top end of wage distribution. We discuss the role
the volume of financial activity plays in the tremendous increase of top financial wages.
Keywords: inequality, wages, finance, France, superstars
JEL classification: D3 microeconomics, distribution; G2 financial insti-
tutions and services; J3 wages, compensation and labor costs
The considerable rise in inequality in the United States during the last 40 years
(Piketty and Saez, 2003) is by now almost common knowledge. Although less
impressive, this trend appears also at an international level, especially in AngloSaxon countries (Atkinson et al., 2011). On the other hand, levels of inequality
in continental Europe and Japan remained much more stable over the last 30
years. Is this contrast due to differences in the type of capitalism in those two
sets of countries (Amable, 2003)—in short, free market capitalism, on the one
hand, and state regulated capitalism, on the other—or is it simply that the same
trend towards greater inequality has been delayed in continental Europe?
Figures from Landais (2008) show that in France inequalities have been increasing again at a substantial rate, but only since the late 1990s.
The analytical description and interpretation of this rise in inequality is only
just beginning. One element of this trend which has been widely commented
on is the tremendous rise in CEO pay over the last 30 years (Bebchuk and
Grinstein, 2005; Gabaix and Landier, 2008; DiPrete and Pittinsky, 2010; Nagel,
2010). Another element is the increase in compensations in the entertainment
industry for sporting or artistic superstars (Rosen, 1981). The social importance
and visibility of these elites, and the availability of their compensation to the
public, may explain part of the focus. However, it is uncertain that they account for a great deal of the rise in inequality. More recently, partly owing to
the financial crisis and the bonus outrage, the importance of financial wages
has come under scrutiny (Kaplan and Rauh, 2010). Philippon and Resheff
(2009) show that in recent years the financial sector has granted wages 50-60%
higher than other sectors for jobs requiring the same level of qualification. Bell
and Van Reenen (2010) estimate that 70% of the recent increase in the share of
the top 1% in the United Kingdom was captured by workers in the financial
industry. Bakija et al. (2010) offer detailed statistics on the occupations of top
earners in the US. According to their dataset, a little more than 30% of the
increase in the share of top earners went to people working in finance.
The goal of the following paper is to investigate the transformation of inequality in France. To that aim, we rely on the DADS data (1976-2007) 1, the French
Social Security wage data for the private sector. Such data enables us to ask
questions about the changing patterns of wage inequality in France. Firstly,
how reliable is the rise in inequality discovered by Landais using self-declared
fiscal sources? If this trend is robust, then who are the beneficiaries? CEO,
managers, experts, entertainment superstars? Since Paris finance is not as
wealthy as that of London or Wall Street, does it nevertheless account for as
much of the rise in inequality?
The paper is organized as follows. In the first section, we will describe the data.
The second section is devoted to the rise in wage inequality over the last 30
years. The third section deals with the changing characteristics of the working
rich in France. In the fourth section we will concentrate on the impact of
finance on the evolution of wage inequality. And finally, in the last section, we
will offer interpretations of the rise of top financial wages.
1. The DADS, a detailed dataset on wages in the private sector
The DADS, Déclaration Annuelle de Données Sociales, is an INSEE (Institut
national de la statistique et des études économiques) statistical dataset based on
an administrative source. In order to collect social contributions for Social
Access to the data was obtained through the CASD dedicated to researchers authorized by
the French Comité du secret statistique.
1
2
Security—payroll taxes, which are more or less proportional to an employee’s
wage—the French Government collects data on all wages from the private
sector. Social contributions from national civil servants are collected through a
different system, and therefore, at present, the latter are not in the database.
On the basis of these administrative records, two main datasets are available.
The first is the Panel DADS (1976-2007), which contains 1/24th of private
sector wage earners from 1976 to 2001 and 1/12th of the same population after
2001. 2 The second dataset is made up of exhaustive files, organized by year and
by region, on all jobs in the private sectors from 1994 to 2007. 3
The great advantage of the DADS is that it offers a very precise image of
wages in France and enables us to calculate fractiles at the very top of the wage
distribution. Moreover, unlike other sources (Philippon and Resheff, 2009;
Kopczuk et al., 2010), wages in the DADS are not top coded. 4 Nevertheless,
there are some obvious limitations in our data that might lead us to both
underestimate and overestimate inequalities in France during recent years.
The notion of wage, as collected in the DADS, is more juridical and fiscal than
economic. It corresponds to the part of the wage on which social contributions
are collected. Two main notions of salary are available: the net salary and the
gross salary.
The gross salary ‘base csg’ is quite exhaustive. It contains not only fixed salary
and variable salary, but also perks (such as car or housing), ‘participation’ and
‘intéressement’—i.e. the two main regulated profit sharing devices (DSDS, 2010,
pp. 35-36). 5 The main limitation is that stock options and free shares are not
counted in this notion of salary, since before 2007, no payroll taxes were
collected directly on these forms of wages. 6 Therefore, we may underestimate
some high salaries like those granted to CEOs of major firms.
Another problem may arise from the fact that the DADS files are organized
according to jobs rather than individuals. Are we to calculate inequalities
among jobs or among individuals? Since workers may have multiple jobs
during the year (successively or simultaneously), especially in an industry such
as entertainment, the latter option appears more relevant. Unfortunately, this
approach is not possible with the exhaustive data files before 2001, as those
files lack individual identification variables. Therefore, before 2001, we limit
They select people born in October every two years until 2001, and every year thereafter.
In the exhaustive files, it is not possible to identify a worker from one year to another or
even, between 1994 and 2001, from one job to another. However, these files contain the
situation in year t and year t-1, so it is possible to measure changes over a two-year period of
time.
4 As outliers possibly resulting from transcription errors may have a significant impact on the
top fractiles, we have excluded salaries that were more than 100 times the P99.99 threshold.
That is, 2 salaries in 1994 over 50 million euros, 1 in 2002 and 4 in 2007 over 100 million
euros.
5 Unfortunately, the DADS do not give any information on the share of the different components of the salary.
6 Dividends and commercial benefits are also not counted in the DADS’ notion of wage. We
therefore underestimate the revenues of CEOs who own an important fraction of their firm.
2
3
3
ourselves to full-time, non-annex jobs 7 and assume that those jobs are held by
different individuals. 8
The notion of hourly wage is not the best approach for studying inequality at
the top of the wage distribution, since we find jobs in consultancy or the
leisure industry where people earn high wages for a very limited set of hours.
Moreover, hours are adjusted by INSEE for what they consider to be extravagant hourly wages. This leans in favour of using yearly wages. Nevertheless,
some workers may have jobs in the private sector for very short periods of
time and therefore appear to be poor on the basis of a yearly wage. In some
cases, they are in fact poor, and that should be accounted for. In other cases,
they might be students, civil servants or self-employed persons who work just a
few hours a year as wage-earners in the private sector. Counting them on the
basis of their yearly wage as low-paid workers would be artificial and lead to an
overestimation of inequality. Moreover, this fraction of the population might
not be stable from one year to the next, which could generate a bias in the
patterns of evolution. In order to avoid this limitation, we restrict our sample,
as in Kopczuk et al. (2010), to salaries that are over half the yearly minimum
wage. 9 We have ensured that moving this minimum threshold did not change
our qualitative results.
Let us summarize: first, in the panel (1976-2007) and in the 2002-2007 exhaustive files, we use the annual sum of gross wages by individuals that are over
half the minimum wage. 10 In the 1994-2001 exhaustive files, we use the annual
gross wage of full-time, non-annex jobs that are over half the minimum wage.
2. The rise in inequality in France
Social scientists generally consider France to be a good example of stability in
inequality during the last 30 years (Atkinson, 2008; Piketty, 2001). We find very
similar results within the scope of our data. The classical P90/P10 ratio drops
by 14% from 4.26 in 1976 to 3.73 in 1984, rises by 10% between 1984 and
1989 and remains very stable through the rest of the period ending at 4.08 in
2007. With this classical indicator, there is no clear sign of a recent surge in
inequality, in contrast to what has happened in other OECD countries. In the
United States, for instance, the same ratio rose continuously by 27% between
1973 and 2000 (Atkinson, 2008, pp. 411-412); in the United Kingdom it
increased by 20% between 1977 and 2000 (Atkinson, 2008, pp. 384-385). Even
Germany, known for the stability of its income distribution, experienced a
sharper increase than France: the P90/P10 ratio increased by 17% between
1989 and 2001.
A job is considered non-annex by INSEE if the compensation is over 3 months of minimum
wage or the number of hours is over 120, the duration over 30 days and the number of hours
per day over 1.5. A job is full time if the number of hours per day is over a certain threshold,
which INSEE calculates for each sector.
8 This approximation first leads us to consider that a person who moves from one job to
another in the middle of the year has two different jobs and is therefore considered two
different individuals. We also exclude individuals who hold many jobs that are annex, part-time
or under the threshold of half the yearly minimum wage. A comparison of the two approaches
is possible for 2001. In the first approach (based on the 2001 files), we analyse inequalities
among 12 670 098 ‘workers’. In the second approach (based on the 2002 files that go back to
2001), our analysis applies to 15 146 231 workers.
9 This restriction is applied to both the panel and the exhaustive files.
10 Before 1999, we use the fiscal gross wage and after 1999, the CSG-based gross wage. As
local civil servants, mail and hospital workers only enter the panel in the 1980s; for the sake of
continuity we also decided to exclude them from the panel. Local civil servants and hospital
civil servants were also excluded from the treatment of the exhaustive files.
7
4
Nevertheless, the P90/P10 ratio may give a biased view of the evolution of
inequality, as it excludes by construction the top wages that went through a
sharp increase in many countries during the last two decades. Therefore, in
order to analyse the evolution of inequality, we calculate fractiles at the top of
the wage distribution following Piketty (2001). As the population panel is very
large (1/24th and 1/12th) and the DADS regional files are exhaustive, there is
no need to compute a Paretian approximation of the threshold or the mean of
each fractile.
We find a global increase of wages, albeit at different rates for each fractile. F090, F90-95, F95-99 and F99-99.9 seem to have increased slowly and regularly
at the rate of +1% a year. F99.9-99.99 and F99.99-100, especially over the last
10 years, have increased more quickly. In 2007, the top 0.01%—that is, the 1
692 highest-paid persons in the private sector earning more than 867 000
euros— were paid on average 1 682 000 euros a year, whereas the F0-90
fractile earned between 7 600 and 46 700 euros in gross salary and on average
22 400 euros a year (Online Table S1, S2, Figure S1).
Therefore, the share of the majority (F0-90) is globally declining, losing 2
points in 30 years. The share of the ‘middle classes’, defined by the fractiles
between P90 and P99.9, remains globally stable or is increasing at a slow rate.
When we move to the top 0.1%, however, we can see a sharp increase of their
share after the year 1996 (Figure 1). The share of the top 0.1% increases by 0.8
points, moving from 1.2% in 1996 up to 2.0%. Half of the 0.8-point increase is
for the top 0.01% and half for the F99.9-99.99.
5
Figure 1 Evolution of the share of the top 0.1% wage earners.
2,0%
1,5%
1,0%
F99.9-99.99
0,5%
F99.99-100
F99.90-100 (panel)
F99.90-99.99 (panel)
F99.99-100 (panel)
F99.90-100 (exhaustive)
F99.90-99.99 (exhaustive)
F99.99-100 (exhaustive)
Note: In 2007, the top 0.1% earners was paid 2.0% of the salaries.
Sources: DADS, panel (1976-2007) and exhaustive job files (1994-2007).
Given that in the panel the share of the top 0.01% is based on a limited number of workers (50-60 up to 2001 and 100-120 after 2001), the robustness of
the measured changes may be questionable. An analysis of the exhaustive files
leads to largely similar results. The top 0.1% increases its share by 0.85%,
moving up from 1.1% in 1996 and 1.95% in 2007. 11 Half of this increase is for
the top 0.01%.
Are the changes that we have described reliable? There are some limitations in
our data, discussed above, which may lead us to both underestimate and
overestimate inequalities. Moreover, INSEE is generally cautious with income
data from DADS, as they suspect that some reporting errors might diminish
the quality of the description of top incomes. Hence, they generally study
lower levels of top incomes (Amar, 2010). INSEE believes that errors have
been diminishing over time (DSDS, 2010). If we consider that the main error
11
0.05 point of this increase seems to be due to the change of definition in 2001.
6
2007
2005
2003
2001
1999
1997
1995
1993
1991
1989
1987
1985
1983
1981
1979
1977
0,0%
1975
Share in total wage bill
F99.90-100
at this level is that of over-reporting, this should lead us to underestimate any
increase in inequality.
Nevertheless, when we compare our trends with those of other sources and
authors like Landais (2008) or Solard (2010), we find similar qualitative results.
Landais, based on income self-declaration, finds that between 1998 and 2006,
the total income of the top 0.01% increased by 64% (capital income and
exercised stock options included), and the wages of the top 0.01% increased by
69%. For the same time period, we find a 123% (exhaustive files) to 131%
(panel) increase in the top 0.01% of wages. Part of the difference may be due
to the fact that Landais works on self-declared net wages and on a larger
population (including civil servants and self-employed persons). Solard finds an
increase in income of 39% for the top 0.01% (capital income and exercised
stock options included) between 2004 and 2007. We find an increase of 44%
of the top 0.01% in the panel and of 36% in the exhaustive files. Although one
is based on wages and the other on full income, the two trends seem to conform to the same pattern.
Moreover, like Kopczuk et al. (2010) for the United States and Landais for
France (2008), we also find with our data that the increase in inequality at the
top of wage distribution during the last 12 years did not correlate with an
increase in wage mobility at that level. Stability in the top 0.1% did slightly
decrease in the first half of the 1980s and increased again in the second half of
the same period, but remained stable in the 1990s and the 2000s. The probability of remaining in the top 0.1% in the following year remains relatively stable,
changing cyclically between 70% and 80%, and after five years it remains
constant between 50-60%. Therefore, despite the randomness of new forms of
remuneration, such as incentive bonuses, this increase in inequality is clearly
not the advent of a lottery society where people suddenly jump to the top or
fall to the bottom of the wage distribution.
3. Changes among the working rich
Empirical studies on inequality (Atkinson et al., 2011, Landais, 2008) usually
discuss several hypotheses in order to explain this trend: biased technological
progress, growth of CEO pay due to the growing size of firms as well as an
increase in superstars’ pay. However, given the limitations of their data, they
are generally unable to provide sufficient empirical evidence to confirm or
infirm either thesis. Although limited to private sector remuneration, the
DADS has two reputable qualities: its historical depth and its economic and
social variables. Thus, it is possible to explore the changes in the social composition of the working rich and to test these hypotheses with this data.
We therefore study the change in the composition of the top 0.1% and the top
0.01%. The panel gives the composition in terms of jobs since 1984, with the
1982 PCS coding. Figure 2 shows some striking transformations within the top
0.1%. The first surprise is the decline of CEOs since 1992. 12 The proportion of
CEOs among the top 0.1% dropped from 50% in 1992 to 22% in 2007. Is this
The increase of the proportion of CEOs between 1984 and 1992 is more difficult to analyse.
In the 1980s, the coding of the PCS was not very reliable, and there were also some errors in
the wages reported. Those two problems make it more likely that middle and lower categories
will be artificially represented in the top 0.1%. This growth may also be due to the change in
the composition of CEO remuneration from capital income to wages. And finally, it is also
possible that the 1980s, a period in which free enterprise and, in particular, small firms were
promoted, was also a time when access to top salaries was obtained mainly through a position
as CEO.
12
7
decline due to a change in the composition of wages and a rise of stock options that are not reported in the DADS? Unfortunately, we lack precise data
on stock options. Several studies converge in showing that stock options
boomed in the 1990s for executives and declined after 2002 (Hamouda, 2010;
Leroy, 2010). 13 Nevertheless, stock options have a decisive impact on executive
pay mainly in big firms. If the moderate and volatile decline of CEOs of large
firms (more than 1 000 workers) within the top 0.1% is artificial, the sharp
decline of CEOs of small firms (less than 1 000 workers), from 45% of the top
0.1% in 1992 to 24% at the end of our period, is less likely to be so.
Although CEO pay for large firms may have risen sharply (Evain, 2007), our
data suggests that the rise in inequality is not due mainly to the traditional elites
directing firms, but rather to lower ranking managers and experts. As long as
the CEOs are not the category that is most responsible for the rise in wage
inequality in France or in the US (Kaplan and Rauh, 2010; Bakija et al., 2010) 14,
the rise of their pay—although higher than that of average salaries (Evain, 2007
; Gabaix and Landier, 2008)—appears to be different than generally analysed.
In most models, CEO pay fluctuates independently of that of other wage
earners. For instance, CEO pay in Gabaix and Landier (2008) is set by an
autonomous market design, whereas in Bebchuk and Fried (2004) it is a
function of executives’ power under the constraint of public outrage. The
pronounced increase in pay among some lower management wage earners
changes our understanding of CEO pay fluctuation, since it might also have
increased the outside options (market model) or have lowered the public
outrage constraint (managers’ power model).
Let us now analyse the impact of lower ranking managers on inequality growth.
Firstly, it must be noted that rising inequality is not due to the rise in the
number of technical professionals such as engineers, whose share stagnates
inside the top 0.1% at a limited level of 8-10%. This element mitigates the
traditional interpretation in terms of biased technological progress. The rise in
inequality does not seem to be due to workers holding the most technical and
scientific knowledge, as was feared in the 1960s and 1970s with the birth of
knowledge and technical societies.
One social category accounts for most of the rise: administrative managers
(‘cadres administratifs’). This group accounted for a little less than 20% in the
mid-1980s. They now represent almost 60% of the top 0.1%. This category
increased by 20 points between 1996 and 2007, a period in which inequalities
escalated once again. Almost half of this increase is due to the category ‘cadres
d’état major’, non-executive chief officers, such as chief financial officers, chief
commercial officers, chief administrative officers etc. Unfortunately, we cannot
go into greater detail, but we suspect, as in the US (Zorn et al., 2005), that the
CFOs, with the ‘financialization’ of firms, are at the root of this trend among
top management. The other half is due to lower ranking managers. We will see
further in the next section whether this pressure on salaries exerted by lower
Proxinvest calculates the Black and Scholes values of shares and stock options granted to
executives for CAC40 since 1998. It rises from 40% to 70% between 1998 and 2001 and drops
back to 45% in the middle 2000s (Leroy, 2010).
14 The comparison with Kaplan and Rauh (2010) is relatively complex, since they predominantly use publicly available information on top executives of publicly traded firms. In their data
we find stability of executives among the top 0.1%. The particularly detailed data from Bakija et
al. (2010) show a decline of the proportion of CEOs among the top 0.1% from 35% in 1997 to
30% in 2005.
13
8
ranking managers is a generalized phenomenon or is due to some limited
sectors of the economy.
Figure 2 Changes in the professional composition of the top 0.1% and
0.01% wage earners.
CEOs
50,0%
40,0%
Non exec. chief
officers
Administrative
managers
30,0%
20,0%
Engineers
10,0%
CEOs (0.1% P)
CEOs (0.01% - E)
Engineers (0.1% - P)
Engineers (0.01% - E)
2007
2005
2003
2001
1999
1997
1995
1993
1991
1989
1987
1985
0,0%
1983
Proportion of persons in top 0.1% and top 0.01%
60,0%
Administrative managers (0.1% - P)
Administrative managers (0.01% - E)
Non exec. chief officers (0.01% E)
Notes: In 2007, 59% of the top 0.01% wage earners were administrative managers. P stands for
panel and E for exhaustive files.
Sources: DADS, panel (1976-2007) and exhaustive job files (1994-2007).
The salaries of sports and media superstars are traditionally under significant
media scrutiny due to the fame of the recipients. Rosen (1981) argues that the
transformation of technology might drive a major income increase for the
most famous superstars, since new technologies such as television, radio, CDs
etc. enable them to replicate their production almost at no cost and become
famous among a wider market. In his survey of the sports economy, Andreff
(2007) also signals the importance of the institutional framework that regulates
both the superstar labour market and the media and advertising industries. In
France, the deregulation of television in the 1980s enabled the multiplication of
TV channels and competition between them for both advertising fees and
broadcasting of superstars. Therefore, superstars could extract a larger share of
the advertising fees. In the early 1990s, the labour market was also deregulated
in the professional sports industry. In European football, the Bosman ruling in
1995 put an end to the limitation on the number of foreign players in European football clubs and, therefore, favoured an increase in transfer fees and
salaries.
As the DADS are a wage database only, it will be difficult to give a complete
picture of the impact of entertainment superstars on inequality. Many artists
9
such as pop singers or writers are paid through copyrights. Nevertheless, we
can at least provide some insight into two categories: sportspeople and film
actors. Sportspeople, like football players, earn their base pay as a salary. And
even if actors are also paid through copyrights and associated rights, a major
part of their income is based on a labour contract and a wage.
Regarding the evolution of the proportion of artists and sportspeople among
the top 0.01%, we must remain cautious in our interpretation, since the detailed 4-digit PCS job code is unreliable before 1997, and rather unreliable
between 1997 and 1999 (with 40-60% of answers either missing or incorrect),
becoming slightly better at the end of the period (missing answers drop from
34% to 18% between 2000 and 2007). Nevertheless, the more aggregate twodigit social categories code does not have such limitations and helps us to see
the global trend.
With all of this in mind, the proportion of artists among the top 0.01% looks
rather stable and is near 2%. 15 Are we missing the real change, since we do not
have their whole income? We do not think so. Newspapers quite often give
rankings of the best-paid actors. In 2007, Le Figaro counted 12 actors over the
threshold of 894 000 euros. 16 In our database, we count 11 actors (PCS=354C)
in rthe top 0.01%. Although their income and expenditure are largely commented on, artists—or at least actors—do not contribute a great deal to the
return of inequality.
The impact of sportspeople seems more sensible. They increase from 4% of
the top 0.01% fractile in the mid-1990s up to 8-10% in the 2000s. In 2007, we
count 112 persons coded 424A professional sportspeople. Although we do not
know their sport, it seems very likely that most of them are football players. 17
Indeed, the transformation of their labour market enabled by the Bosman
ruling seems to have had important effects on wages in the sports industry.
In the end, however, although we find that superstars, or at least football
players, do have an effect on inequality, the effect remains limited compared to
the rise in salaries of a fraction of business managers that we will try to define
more precisely in the next section.
4. The impact of finance on the resurgence of inequality
A sector approach enables us to describe more precisely the type of business
managers that contributed the most to the increase in inequality. It is also a
way to address the question of the impact of finance, an industry under scrutiny since the subprime crash and the ensuing bonus outrage.
Some sectors such as manufacturing, retail and restaurants, transport and
communication are now less represented at the top of the wage hierarchy than
they were 30 years ago. For instance, 38% of the top 0.1% worked in manufacturing in 1976, whereas only 14% did so in 2007. On the other hand, service to
business, finance and, to a lesser extent, entertainment and other services
increased their share among the highest paid workers. In 1976, 10% of the top
0.1% were in service to business, and 6% were in finance. In 2007, these
figures were 26% and 24%, respectively (Online Figure S2).
The data show a significant discontinuity in 2001 due to the fact that before this date we
cannot sum multiple jobs.
16 Source: ‘Le palmarès 2008 des acteurs’, Le Figaro, February 22, 2008.
17 We find several football clubs among the firms paying the highest salaries. Moreover, there
were not so many international superstars in cycling or tennis during the period, and other
sports like basketball or rugby pay much less in France.
15
10
At first glance, finance still seems to lag behind service to business among the
top 0.1%. However, the increase and decrease in the different sectors at the
top should be compared to their evolution as a whole inside the private sector.
Thus, service to business is a sector in which the headcount has grown quite
rapidly during the last quarter of a century, whereas the number of workers in
finance has remained a fairly stable proportion of the private sector. 18 We
therefore compute the odds ratio of (a) the percentage within the top 0.1%
with (b) the percentage within the rest of the French private sector, in order to
control for the fluctuations in the size of the sectors among the global population. The result is very striking: in the finance industry of the early 1980s,
financial workers had twice the presence in the top 0.1% as they had under this
threshold. This ratio increased smoothly in the 1980s and very sharply after
1995. In 2001, the ratio peaked at 10, as a result of the considerable bonuses
granted after the excellent market year of 2000. The 2001-2002 crisis lowered it
to 7, and the following boom pushed the ratio back to 10 (Online Figure S3).
Although some sectors might be over-represented among the top salaries, like
service to business or entertainment, no overrepresentation is as considerable
as that achieved by the finance industry in the last 10 years.
We find a correlation between the notable rise in the overrepresentation of
finance among the top 0.1% after 1995 and the rise in inequality in the same
period. Therefore, we can try to quantify the contribution of this sector to this
increase, following Bell and Van Reenen (2010). We calculate the contribution
of finance, service to business, entertainment and other sectors to the 0.85point increase of the wage share. We find that finance contributed to 48% of
this rise, whereas service to business and other sectors each contributed nearly
23%, and entertainment to 8% of the rise (Figure 3).
2.8% of the private-sector workforce was working in finance at the end of the 1970s. This
proportion rose to 3.5% in the mid-1980s, declining to 2.9% in 2000 and stabilizing around 3%
thereafter (Panel).
18
11
Figure 3 Increase of the top 0.1% wage earners’s share of all wages by professional sector.
0,90%
All
0,80%
0,60%
0,50%
Finance
0,40%
0,30%
Service to business
0,20%
Other sectors
0,10%
Entertainment
All
Finance
Service to business
Other sectors
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
-0,10%
1995
0,00%
1994
Change in share of overall wages
0,70%
Entertainment
Notes: Between 1996 and 2007, the share of the top 0.1% wage earners globally increased by
0.85 points, and the share of finance within this fractile increased by 0.40 points. We have
corrected the economic activity for holdings (see online appendices, sector coding).
Sources: DADS, exhaustive job files (1994-2007).
When we move into the top 0.01%, we find that finance makes a contribution
of 57% to the increase in the share of the working rich (Table 1). At the end of
the period, finance constitutes 37% of the headcount of the top 0.01%, which
are 19.4 times more present at this level than below. Overrepresentation of this
sector within the top fractile is much higher than that of service to business
(2.3) or entertainment (6.7). Moreover, we must not forget that we have a small
discontinuity in 2001 in our series of exhaustive files that may lead us to
overestimate the increase between 1996 and 2007 and to underestimate the
impact of finance on this increase. When we look at calculations on the basis
of the panel data, finance makes a greater contribution to the increase in the
top fractiles – between 47% and 70% (Table 1).
12
Table 1 Contribution of finance to the increase in the share of the top fractiles
Top 10%
France panel
UK 1998-2008
US 1997-2005
Top 0.1%
Top 0.01%
Share in 1996
26.45%
5.74%
1.20%
0.27%
Share in 2007
27.74%
7.06%
2.01%
0.65%
1.29%
1.32%
0.81%
0.38%
51%
47%
57%
69%
Share in 1996
25.67%
5.43%
1.10%
0.23%
Share in 2007
27. 70%
6.97%
1.95%
0.60%
Increase in the share
Contribution of finance
to this increase
2.03%
1.54%
0.85%
0.38%
33%
39%
48%
57%
Increase in the share
Contribution of finance
to this increase
3.00%
1.80% -
-
73%
72% -
-
Increase in the share
Contribution of finance
to this increase
France exhaustive files
Top 1%
Increase in the share
Contribution of finance
to this increase
2.54%
1.65%
32%
31%
Notes: Between 1996 and 2007, according to the panel, the share of the top 10% globally
increased by 1.29 points, and the share of finance within this fractile contributed 51% to this
increase.
Sources: France: DADS, panel (1976-2007) and exhaustive job files (1994-2007). UK: Bell and
Van Reenen, 2010, table 3 – ASHE. US: Bakija et al., 2010, Table 5 & 6.
Our figures for France are in between those that can be calculated from Bakija
et al. (2010) for the United States and those found by Bell and Van Reenen
(2010) for the United Kingdom. In all three countries, finance played a major
role in the return of wage inequality, contributing to a third (US), a half
(France) and three quarters (UK) of the rise of top wages. This strong contribution must be balanced against the much more limited share of finance within
the workforce at the end of the period: 2% in France, 3% in the United Kingdom and 3.4% in the United States. 19 Beyond this overrepresentation of
finance at the top and in the surge, we nevertheless find some striking differences. Part of the discrepancy is due to differences in data, methods and
industry definition. For instance Bakija et al. (2010) analyse the full income. For
France, the limitation to private sector wages only (excluding therefore many
self-employed professionals such as doctors and lawyers) and to cash salaries
(excluding CEO stock-options and shares) leads us to recognize that our
estimation of the contribution of finance to the surge in inequality is more
likely an upper bound. Nevertheless, part of the divergence also seems to
come from the nature of the phenomenon. 20 In the United States, growth of
top incomes seems to additionally concern many other sectors and professions,
such as lawyers, real estate professionals and non-finance business executives
and managers. In France and in the United Kingdom, where the surge in
inequality is much more recent, it is more concentrated in finance, a sector that
could be viewed as the avant-garde of this new trend. The difference in the size
We use the 2007 Emploi survey for France (civil servants and self-employed persons
included), the 2007 Labour Force Survey for the United Kingdom and the 2008 County Business
Patterns for the United States.
20 Bell and Van Reenen (2010) also estimate the contribution of finance to the growth of top
incomes at 60%, a figure that can be more easily compared to that of Bakija et al. (2010).
19
13
of the two financial centres, Paris and London, probably accounts for this
difference between France and United Kingdom. 21
Finance, therefore, appears to have played a major role in the return of wage
inequality in France. How did this trend arise? Has remuneration in finance
been growing at all levels compared with the rest of the economy? Or is the
deviation due to certain levels of income distribution?
In order to analyse the structure of the premium for the financial sector, we
run annual cross-section wage quantile regressions with the following control
variables: sex, age, square age, seniority, square seniority, social group (managers, technicians, clerks, workers), geographical location (Paris region versus rest
of France), number and square number of employees in the firm and a variable
for the financial sector (Online Figure S4). In 1976, the premium for the
financial sector was 23%. It went down to 11% in 1989 and climbed back to
22% in 2007. Quantile regressions show a similar evolution for most of the
thresholds except for P99, which has been increasing in the medium term from
9% to 37%. The hierarchy of premium also changed substantially. At the end
of the 1970s, the premium was larger at the bottom of the distribution (ranging
from 9% for P99 to 34% for P10); in the 2000s, apart from the P99, most of
the thresholds converged between 19 and 23%. Therefore, the contribution of
finance to the increase in wage inequality mainly seems to be linked to the
evolution of its top 1% and to the considerable rise in inequality within this
sector. Hence, in less than 12 years, the share of the top 1% within finance
moved from 6% to 12% of the wage share.
Who is responsible for the increase in inequality within the finance industry?
Following Kaplan and Rauh (2010), we would expect the employees who are
most tied to the financial markets to be linked to this phenomenon. In 2003,
INSEE reformed its PCS code and introduced a new category, financial market
managers (cadres des marchés financiers), among whom we find traders, salespeople, financial analysts, portfolio managers, brokers, financial engineers and risk
managers. The category reflects quite well what people in the market generally
call ‘front offices’. This group is very likely to capture the impact of the growth
of financial markets on wages. Unfortunately, there are a few drawbacks. First,
the category does not allow close scrutiny of the 12 years under consideration
and does not allow us to view the great boom of the financial markets during
the second half of the 1990s. Second, due to its novelty, firms might not be
accustomed the new code for people that were traditionally coded as bank
managers (cadres de banque). Third, we do not know if heads of trading rooms
and heads of desks, the highest-paid employees on the financial markets, are
always coded as such. Despite its limitations, the category is a good proxy for
the recent impact of the financial market (with perhaps a little underestimation
of the actual scope).
During the last five years, the importance of this category grew in the top
fractiles of the financial sector. They made up 20.6% of the top 1% in finance.
They represented 27.8% in 2007. The same growing trend is observable within
the French private sector. By 2007, at the end of the period studied, financial
market managers accounted for 13% of the top 0.01%—that is more than
In 2007, 3.8% of the workforce worked in finance in the Paris region (Emploi survey), while
5.5% did the same in the London region (Labour Force Survey). It should be noted that when we
do the same decomposition on the Paris region only, we find that the rate of increase in
inequality and the contribution of finance to this increase are more similar to the situation in
the United Kingdom. The increase of top 0.1% share in Île de France is 1.4% and the contribution of finance amounts to 60% of that share.
21
14
professional sportspeople—and were 150 times better represented than in the
rest of society. Therefore, although we do not have much historical depth, the
impact of market managers on the 2003-2007 rise in inequality suggests that it
is mainly the boom of financial market activity since the mid-1990s that fuelled
inequality in finance.
5. Elements of interpretation
Finally, Figure 4, which compares the evolution of top salaries, allows us to
sum up some of our main findings. In the figure, we analyse the evolution of
the top 100 finance managers (people working in finance sector as ‘cadres’),
the top 100 non-finance and non-entertainment managers, the top 100 CEOs,
the top 25 sportspersons and the top 20 wage earners in the movie, TV and
video sectors (most of whom are actors). Between 1996 and 2007, wages
increased by 1.5 in this latter group, by 3.3 in sports and among the top CEOs,
by 3.6 among the top non-finance managers and by 8.7 among the top 100
finance managers. On the basis of salary comparison, top finance managers
clearly surpass other elites both by the pace of growth and by the level of pay
at the end of the period. We nevertheless must remain cautious, as we lack
information on other forms of compensation such as shares or stock options.
In order to partly overcome this limitation, we have estimated for top CEOs
the probable changes in compensation, stock options included, by applying the
share of stock options to them that Proxinvest (2009) calculated for CAC40
executive teams. 22 The pace of increase of top finance managers pay (salary
only) was double that of top CEOs, and the former almost caught the level of
the latter in 2007. Moreover, although remuneration in shares and above all in
stock options was not particularly common before the 2008 crisis, it is likely
that this small financial elite did increasingly benefit from this form of pay. 23
We must note that the top100 CEOs and the executive teams of CAC40 (approximately
400-500 executives) is not the same population. The best-paid CEOs in cash may be outside
CAC40, or even outside the SBF 250 spectrum. They may also work in non-public firms.
Nevertheless, there is enough proximity between the two populations to use as a rule of thumb
the Proxinvest series, which has to my knowledge the most historical depth. Proxinvest
calculates the Black and Scholes value of options granted to CEOs, with a small discount for
the invalidity period. The Black and Scholes value of stock options is a good representation of
compensation and, thus, of inequality in the firm from the point of view of the shareholder. It
may be more questionable in order to estimate the inequality of standard of living, since
options may or may not be exercised.
23 In an internet survey launched with efinancialcareers.fr in September 2008—a snow ball
sample (n = 992), broadly representative of the diversity of the financial industry (with a junior
bias)—we found that the proportion of persons earning stock options was only 3%, and the
proportion of those with shares was 4%. Since the subprime crisis, pay regulation policies have
recommended that firms pay roughly one-third of bonuses in restricted shares. Hedge fund
managers who are partners of their firm may also earn directly commercial benefits in addition
to fixed salary, bonuses and stock options. However hedge funds remain rare and small due to
restrictive legislation in France.
22
15
Figure 4 Evolution of the top wages for several well-known jobs.
10 000 000
4 652 388
871 411
1 042 977
2007 constant euros
1 000 000
535 257
Top 100 finance managers
Top 25 sportspersons
Top 100 non-finance managers
Top 20 movie sector
Top 100 CEOs
Top 100 CEOs (stock estimated)
Notes: In 2007, the top 100 finance managers were paid 4,652,388 euros on average per year.
Yearly figures have been calculated in 2007 constant euros.
Sources: France – exhaustive job files DADS (1994-2007) and Proxinvest (2009).
Therefore, the most scrutinized, highly paid professionals, such as CEOs and
entertainment superstars, are not responsible for most of the increase in
inequality in comparison with finance managers, in particular heads of desks
and heads of trading rooms.
Several interpretations have been provided in order to explain this extraordinary wage trend in the financial industry. The importance of human capital has
been researched both by Philippon and Resheff (2009) for the US, and also for
one of France’s main banks (Godechot, 2011). Despite the importance of
higher education degrees at the core of the financial markets, even very detailed education variables in traditional wage equations fail to explain the wage
structure or its evolution.
A great deal of recent research links the way in which compensation in the
financial industry has evolved with the evolution of the volume of activity
(Meunier, 2007; Kaplan and Rauh, 2010; Célérier, 2010). Kaplan and Rauh
(2010) outline an impressive series on the rise in the amount under management in hedge funds, increasing from 20 billion in 1986 to 1 trillion in 2004.
Although the volume of shares exchanged on the Paris stock market (Figure 5)
may not be fully representative of the increase in the volume of financial
activity—missing over-the-counter, fixed-income or foreign financial products—it is at first glance a reasonably good approximation of investment bank
activity, which, in the end, in France is mainly an activity of intermediation
(brokering, equity derivatives pricing and marketing etc.). Figure 5 clearly
shows how financial activity boomed at a very rapid rate during three periods:
1984-1987 (+70% per year), 1995-2000 (+50% per year) and 2004-2007 (+
25% per year).
16
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
100 000
If we compare the evolution of the top 100 finance managers during the last 12
years with our volume index, results are strikingly congruent with the idea that
there is a strong link between volume and compensation 24. Between 1995 and
2006, volume was multiplied by 8.85. During the same period, the top 100
finance managers’ wages were multiplied by 8.26. The correlation between the
two curves (r = 0.92), although not perfect, is nevertheless impressive. 25
Although we must remain cautious and acknowledge that a correlation based
on 14 observations may be spurious, let us note that this relation is also supported by more qualitative elements collected in fieldwork research. If banks
tended in the 1980s and the 1990s to replace formulas for bonus by qualitative
discretionary pay, they also introduced some collective formulas based on net
revenue in order to generate bonus pools at the departmental level (for instance, equity derivative departments) (Eccles and Crane, 1988). Those formulas (more generally a simple proportion of the net income) are volume driven,
and in practice heads of departments’ share of the bonus pool did not seem to
diminish with growth of the headcount under their supervision.
Figure 5 Value of shares exchanged on the Parisian stock market and the top
100 finance managers’ wages.
10 000 000
1 712 090
4 652 387 €
1 000 000
193 395
535 257 €
100 000
10 000
1970
1975
1980
1985
1990
1995
2000
2005
Value of shares (Constant 2007 million Euros)
Top 100 finance managers' wages *
Notes: In 2006, 1,712,090 million euros’ worth of shares were exchanged on the Paris market.
In 2006 (that is, in 2007 for their 2006 performance), the top 100 finance managers were
granted 4,652,387 euros on average. *We rescale the top 100 finance managers curve a) in year
n-1, as bonuses are generally paid in year n for the year n-1 performance, b) so that the two
curves share the same reference point: 1995.
24 The effect of the first boom of the 1980s on compensations is more difficult to detect for
the following reasons: financial markets represented only a small fraction of the finance
activities at that time, the panel at 1/24th of the population lacks precision, and a large part of
these activities were then carried out by the Agents de change, traditional French brokers and
their employees, who were largely paid via heterodox means.
25 If we regress the logarithm of the top 100 average wages on the logarithm of volume index,
we find an R2 of 85% and a very significant coefficient of 0.9.
17
2010
Sources: France – exhaustive job files DADS (1994-2007) and Euronext, Euronext Fact Book,
Historical Series (Turnover), accessed at http://www.euronext.com/editorial/wide/editorial20786-EN.html.
The traditional but intriguing correlation between CEO pay and firm size was
recently given an explanation that could be relevant for financial market
managers. Gabaix and Landier (2008) explain how the heterogeneity of CEO
talent may be multiplied by a skewed distribution of volume. They develop a
model where the biggest firm hires the best CEO in order to maximize the
impact of the CEO for its shareholders. In this model, the best CEO does not
need to be a ‘superhero’, but only to be slightly better than the 250th CEO
(that is, to increase the capitalization by 0.016%) in order to get, due to the
skewness of the distribution of company size, a multiple of its salary (for
instance 5 times more in their calibration). In their model, while the crosssection relation between logarithm of pay and logarithm of size is only 1/3, it
increases to 1 in a longitudinal approach.
This mechanism was also invoked for financial labour markets by several
authors (Meunier, 2007; Kaplan and Rauh, 2010), and Célérier (2010) developed a model based partially on this idea. In the same spirit, if a star trader can
get 5.1% return on equity instead of 5.0% as an ordinary trader, he will be
matched to the biggest portfolio and will get an extra bonus of 0.1% of the size
of the portfolio (for instance, 1 million euros more if he is matched to a 1billion-euro portfolio). If we follow this perfect market mechanism of matching of size and talent, which requires the following two strong conditions,
perfect mobility and perfect knowledge of the hierarchy of talent, the hierarchy
of pay only follows a natural, independent hierarchy of talent.
Two cases may be distinguished depending on the generality or the specificity
of talent for finance.
If talent is general (like in Gabaix and Landier’s main framework), pay will be
distorted by the skewness of the distribution of volume. But we still cannot
talk of rents here. The theoretical result rests on perfect mobility both within
and between sectors. This first hypothesis is at odds with the data: the mobility
rate from non-finance to finance did not increase during the 1996-2001 finance
boom, although the increase in size of projects in finance should have attracted
more non-finance top performers. 26
If talent is sector specific, the model still holds under the hypothesis of perfect
mobility within the sector. But if a boom in finance increases volume and
therefore pay, we may be allowed to talk of a global rent in finance, in the
sense that after the boom there will be an excess in earnings over the amount
necessary to keep the factor in its current occupation (Shepherd, 1970).
Moreover, the ‘natural’ origin of talent that plays an important role in these
models in order to justify these compensations may be questioned. Thus, Oyer
(2008) shows that MBAs will be all the more likely to work in finance, to stay
in the sector long-term and to earn top wages if they have graduated in a bull
market. This statement may seem quite trivial, but it shows clearly that finance
managers ‘are largely “made” by circumstance rather than “born” to work on
Wall Street’ (Oyer, 2008). Were finance talent natural, there would be no
reason for talented finance people to be more numerous in bull years than in
bear years. And were they incorrectly selected in bull years, we should see more
In 2001, only 1.1% of the top 10% of managers in 1996 working in non-finance had moved
to finance (booming period). In 1996, 1.7% of 1991’s similar population did so (non-booming
period).
26
18
of them leaving finance, which is clearly not the case. Oyer concludes that
MBAs develop finance-specific human capital shortly after taking jobs on Wall
Street.
Therefore, we can retain the idea that the volume of financial activity is responsible for the increase in pay without linking it to a natural and intangible
hierarchy of talents. Talent may not only be natural, observed and eventually
revealed by the financial industry (Célérier, 2010), but also acquired on the
job 27, a process which has been, therefore, largely funded by the employer.
Moreover, talent may not be the only capacity an employee might acquire on
the job. Other research has developed a model of financial activities where
financial operatives appropriate the key assets of the firm and can threaten to
move those assets to a competitor in the same sector (Godechot, 2008). Those
assets can be traditional human sector-specific capital such as knowledge and
know-how, but we must also include more material assets such as software and
databases, as well as social capital such as customer relations or productive
teams. In our 2008 paper, we analyse in detail a case of ‘hold-up’. In a 2001
wage renegotiation, the head of a trading room and his deputy were granted 10
and 7 million euros, respectively, by effectively threatening to move their
whole teams, and therefore the core of the firm’s financial activity, to a competitor. Although those two individuals might have been very talented, what
was at stake in this wage renegotiation was not their initial talent but their onthe-job accumulated social capital that enabled them to expropriate part of the
firm’s assets. Thus, the specificity of finance may not be its greater sensitivity
to talent (Célérier, 2010), but rather the fact that physical property rights,
intellectual property rights such as patents and labor contract devices like noncompete clauses are much less effective at protecting the firms’ assets against
worker appropriation.
Therefore, in such a model, if the accumulation of movable assets allows a
financial worker to capture a fraction of financial activity, the growth of the
latter leads to a growth in his pay. Even if we cannot totally rule out that their
remuneration can be explained in a superstar framework, we can, however,
find an explanation for the trend in finance pay without considering that the
financial elites are the natural elites of society. Nevertheless, we must also
recognize one limit of the model: although it explains the share of the joint
value added between financial operatives and firms, it does not give an explanation of the growth of volumes in finance or of the value added captured by
this sector.
6. Conclusion
France has experienced a strong increase in inequality over the last 12 years.
Half of the increase of the share of the top 0.1% is due to an increase in pay
among top finance managers. On the other hand, CEOs and entertainment
superstars did not seem to play a major role in the increase in inequality.
The interpretation of this trend is only just beginning. We nevertheless find a
striking correlation between the top 100 finance managers’ pay and turnover
on the Paris stock market. The relationship between the volume of financial
activity and pay may not only be due to a multiplicative effect of volume on
Célérier builds a model where finance is a sector more sensitive to talent. Talent is discovered after the first working period. In the model, this talent can either be an initial talent that is
revealed or on-the-job acquired talent. Nevertheless, in her argumentation, Célérier favours the
first hypothesis.
27
19
initial talent, but also to the fact that workers in finance can appropriate a share
of the firm’s assets, assets which have been growing rapidly over the last 12
years.
Although the basic models linking volume of financial activity and pay might
be relatively similar, should they be based on initial talent or on acquired assets,
more work is needed in order to separate the contribution of these two factors.
This research program has an obvious policy implication. With the 2008
financial crisis, some social analysts pleaded in favour of a tax on financial
wages, and the UK and France have experimented with this tax for a limited
time. In this debate, taxing talent or taxing rents does not have the same
political significance.
It should also be noted that the taxation of finance workers and the taxation of
high incomes has received contradictory attention in the public debate. France,
during the last decade, as in many developed countries, has been lowering the
tax rates for the highest incomes, after some consideration of the positive
effects of those elites on overall activity. At the same time, CEOs during the
whole decade, finance workers after 2007 and sportspeople after the 2010
World Cup defeat, have been widely criticized. Both the meritocratic character
of their pay and the usefulness of their economic role have been subject to
debate. It should be noted that these categories are not marginal among top
wages in France. In the top 0.01% of wages for 2007, we find nearly 40% of
finance workers, 20% of CEOs and 10% of sportspeople. Taxing this fractile
of salary more would be another way (perhaps more easily achieved than a
sectorial tax) to redistribute those salaries, which more and more citizens
consider as rents.
20
References
Amable, B. (2003) The diversity of modern capitalism, Oxford, Oxford University
Press.
Amar, M. (2010, Avril) Les très hauts salaires du secteur privé, insee première, INSEE
Première No. 1288, Paris, Institut national de la statistique et des études
économiques,
accessed
at
http://www.insee.fr/fr/ffc/ipweb/ip1288/ip1288.pdf on February 8, 2012..
Andreff, W. (2007) 'Régulation et institutions en économie du sport', Revue de la
régulation, 1, accessed at http://regulation.revues.org/1274 on February 1, 2012.
Atkinson, A.B. (2008) The Changing Distribution of Earnings in OECD
Countries, Oxford, Oxford University Press.
Atkinson, A.B., Piketty, T. and Saez, E. (2011) ‘Top Incomes in the Long Run
of History’, Journal of Economic Literature, 49, 3-71.
Bakija, J., Cole, A., Heim, B. (2010) Jobs and Income Growth of Top Earners and the
Causes of Changing Income Inequality: Evidence from U.S. Tax Return Data, Department of Economics Working Papers, Williamstown, Williams College, accessed
at
http://web.williams.edu/Economics/wp/BakijaColeHeimJobsIncomeGrowth
TopEarners.pdf on February 1, 2012.
Bebchuk, L.A. and Fried, J. (2004) Pay without Performance: The Unfulfilled Promise
of Executive Compensation, Cambridge, MA, Harvard University Press.
Bebchuk, L.A. and Grinstein, Y. (2005) ‘The Growth of Executive Pay’, Oxford
Review of Economic Policy, 21, 283–303.
Bell, B. and Van Reenen, J. (2010, April) Bankers’ pay and extreme wage inequality
in the UK, CEP Special Paper No. 21, London, Centre for Economic Performance,
London
School
of
Economics,
accessed
at
http://eprints.lse.ac.uk/28780/1/cepsp21.pdf on February 8, 2012.
Célérier, C. (2010) ‘Compensation in the Financial Sector: Are all Bankers
Superstars?’, paper presented at Paris School of Economics, Lunch Seminar,
Paris,
January
18,
accessed
at
http://www.efmaefm.org/0EFMAMEETINGS/EFMA%20ANNUAL%20M
EETINGS/2011-Braga/papers/0485_update.pdf on February 28, 2012.
DiPrete, T. and Pittinsky, E. (2010) ‘Compensation Benchmarking, Leapfrogs,
and the Surge in Executive Pay’, American Journal of Sociology, 115, 1671-1712.
DSDS (2010) DADS, Guide méthodologique, Validité 2008, Dijon, INSEE.
Eccles, R. and Crane, D. (1988) Doing deals: investment banks at work, Cambridge,
MA, Harvard University Press.
Evain, F. (2007) Le salaire des chefs d’entreprises, moyennes et grandes, INSEE Première No. 1150, Paris, Institut national de la statistique et des études
économiques.
Gabaix, X. and Landier, A., (2008) ‘Why has CEO Pay increased so much?’,
Quarterly Journal of Economics, 123, 49-100.
Godechot, O. (2008) ‘“Hold-up” in finance: the conditions of possibility for
high bonuses in the financial industry’, Revue française de sociologie (Supplement
annual English edition), 49, 95-123.
21
Godechot, O. (2011) ‘Le capital humain et les incitations sont-ils les deux
mamelles des salaires dans la finance ?’, Revue d’économie financière, 104, 145-164.
Hamouda, M. (2010) ‘Le bilan des stock-options en France de 1997 à 2003’,
Management et avenir, 45, 149-167.
Kaplan, S. and Rauh, J. (2010) ‘Wall Street and Main Street: What Contributes
to the Rise in the Highest Incomes?’, Review of Financial Studies, 23, 1004-1050.
Kopczuk, W., Saez, E. and Song, J. (2010) ‘Earnings Inequality and Mobility in
the United States: Evidence from Social Security Data since 1937’, Quarterly
Journal of Economics, 125, 91-128.
Landais, C. (2008) Top incomes in France (1998-2006): Booming inequalities?, PSE
Working Paper, Paris, Paris School of Economics, accessed at
http://www.stanford.edu/~landais/cgi-bin/Articles/topincomes.pdf
on
January 11, 2012.
Leroy, P.-H. (2010) ‘La rémunération des dirigeants et des traders dans le
secteur financier’. In Rapport moral sur l’argent dans le monde 2010, Paris, Association d’économie financière, pp. 181-188.
Meunier, F. (2007) ‘Les hauts salaires dans la banque : talent ou collusion ?’,
Revue française d’économie, 22, 49-71.
Nagel, G. L. (2010) ‘The Effect of Labor Market Demand on U.S. CEO Pay
Since 1980’, The Financial Review, 45, 931-950.
Oyer, P. (2008) ‘The Making of an Investment Banker: Stock Market Shocks,
Career Choice, and Lifetime Income’, The Journal of Finance, 63, 2601–2628.
Philippon, T. and Reshef, A. (2009) Human Capital in the U.S. Financial Industry:
1909-2006, NBER Working Paper No. 14644, Cambridge, MA, National
Bureau
of
Economic
Research,
accessed
at
http://www.nber.org/papers/w14644.pdf on January 11, 2012.
Piketty, T. (2001) Les hauts revenus en France au XXe siècle, Paris, Fayard.
Piketty, T. and Saez, E. (2003) ‘Income Inequality in the United States, 19131998’, Quarterly Journal of Economics, 118, 1-39.
Proxinvest (2009) La rémunération des dirigeants des sociétés cotées, Paris,
Proxinvest.
Rosen, S., (1981) ‘The Economics of Superstars’, American Economic Review, 71,
845-858.
Shepherd, R. (1970) ‘Economic Rent and the Industry Supply Curve’, Southern
Economic Journal, 37, 209–211.
Solard, J. (2010) ‘Les très hauts revenus: des différences de plus en plus marquées entre 2004 et 2007’. In Les revenus et patrimoine des ménages, édition 2010,
Paris, Insee, pp. 45-64.
Zorn, D., Dobbin, F., Dierkes, J. and Kwok, M. S. (2005) ‘Managing Investors:
How Financial Markets Reshaped the American Firm’. In Knorr Cetina, K.
and Preda A. (eds) The Sociology of Financial Markets, Oxford, Oxford University
Press, pp. 269-289.
22
SER Supplementary Data on line
Is finance responsible for the rise in wage inequality
in France?
Supplementary Data on line
Figure S1. Evolution of constant wages of the different fractiles (in
euros, 2007)
Note: In 2007, the mean salary in the top 0.01% was 1,682,324 euros.
Sources: DADS Panel (1976-2007) and exhaustive job files (1994-2007).
23
Figure S2. Finance and other sectors in the top 0.1%
45.0%
Finance
Insurance
40.0%
Entertainment
Manufacturing
Construction
30.0%
Retail and restaurants
Transport & Com.
25.0%
Other services
20.0%
State, education, health
15.0%
10.0%
5.0%
2007
2005
2003
2001
1999
1997
1995
1993
1991
1989
1987
1985
1983
1981
1979
1977
0.0%
1975
Proportion of employees in top 0.1%
Service to business
35.0%
Notes: In 2007, 26% of the top 0.1% worked in service to business. We correct the economic
activity for holdings coded among the service to business sector.
Source: DADS, Panel (1976-2007).
24
Figure S3. Overrepresentation within the top 0.1%
Notes: In 2007, there are 10.3 times (in terms of odds ratio) more finance employees in the top
0.1% than there are in the rest of the distribution. We correct the economic activity for
holdings (cf. Appendices, sector coding).
Source: Panel DADS (1976-2007).
25
Figure S4. Evolution of Finance Premium.
Notes: On the figure is plotted the wage premium for finance estimated through both OLS
regressions and quantile regressions. Working in finance in 1976, increases by 34% the P10
threshold.
Source: DADS, Panel (1976-2007).
26
Sector coding
Sectors
Finance
Insurance
Entertainment
Service to business
Industry (and agriculture)
Construction
Retail and restaurants
Transport and communications
Other services
State, education, health
NAP
89, 7801
88, 7802
86
76, 77
01-54, 56
55
57-67
68-75
90-97, 99
98
NAF 1993
65, 67.1
66, 67.2
92
74
01-41
45
50-57
60-64
70-73, 90-99
75-85
Correction of sectors for holdings
With the financialization of the firm, heads of firms are often constituted as
holdings, managing many different units involved in many different economic
activities. Their economic sector is difficult to code unilaterally. Therefore,
before 1993, in the NAP nomenclature, INSEE gave them their own division
(76). After 1993, in the NAF nomenclature, we find them inside the service to
business division (74), the “administration of firm” code, “741J”, with other
activities of firm management or representation. Holdings, therefore, are not
totally isolated: in the Panel, we count 1756 individuals working for holdings in
1992 whereas 7995 are working in the 741J “administration of firm” code in
1993.
Heads of groups, where we generally find the highest salaries, working in
industry, retail, construction, transport, and finance, will therefore be coded in
service to business. This type of coding might overestimate the role of service
to business in higher fractiles. In order to eliminate this bias we tried to correct
the coding. We used the 2002-2007 Lifi survey in order to correct the sector
for heads of groups. When a head of group is coded as a holding we assign to
it the sector of its biggest (in head-count) subsidy. For the years before 2002,
we use the 2002 Lifi survey. The approximation is not too bad, as far as during
the period holdings are generally created rather than destroyed. We reassign
20% of workers coded in 741J in 2007, 16% in 2002, 13% in 1995, and 30% of
workers coded in NAP76 in 1991, 20% in 1976.
Within the 2007 top 0.1% fractile, this correction helps to reduce the proportion of wage-earners in service to business from 31% to 26%, and to increase
that of industry from 11% to 14%, and that of retail and restaurants from 9 to
10%. It does not have much impact on other sectors, especially on finance.
27
Table S1. Thresholds, means and standard estimates of yearly gross wages for different fractiles of the distribution (Panel). Constant 2007 euros.
Year
1976
1977
1978
1979
1980
1982
1984
1985
1986
1987
1988
1989
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
Selection
threshold
4,528 €
4,851 €
5,234 €
5,314 €
5,412 €
5,805 €
6,026 €
6,157 €
6,253 €
6,307 €
6,311 €
6,344 €
6,504 €
6,604 €
6,637 €
6,671 €
6,757 €
6,864 €
6,979 €
7,138 €
7,218 €
7,256 €
7,396 €
7,494 €
7,626 €
7,884 €
7,746 €
7,524 €
7,603 €
F00-90
Mean
F00-90
Std
P90
F90-95
Mean
F90-95
Std
16,049 €
16,573 €
17,305 €
17,482 €
17,716 €
18,066 €
18,366 €
18,638 €
18,961 €
19,089 €
19,079 €
19,359 €
19,919 €
20,131 €
20,286 €
20,139 €
20,319 €
20,301 €
20,324 €
20,471 €
20,247 €
20,287 €
20,655 €
21,011 €
21,063 €
21,205 €
21,834 €
21,960 €
22,314 €
6,544 €
6,550 €
6,729 €
6,721 €
6,827 €
6,889 €
6,934 €
7,072 €
7,269 €
7,394 €
7,524 €
7,725 €
7,940 €
8,004 €
8,172 €
8,255 €
8,202 €
8,135 €
8,149 €
8,157 €
8,184 €
8,196 €
8,324 €
8,383 €
8,403 €
8,365 €
8,680 €
8,769 €
8,898 €
32,673 €
33,243 €
34,487 €
34,543 €
34,964 €
35,752 €
36,396 €
37,181 €
38,123 €
38,629 €
38,905 €
39,774 €
40,896 €
41,173 €
42,011 €
41,908 €
41,997 €
41,688 €
41,750 €
42,016 €
42,531 €
42,767 €
43,691 €
43,937 €
43,920 €
44,103 €
45,465 €
45,531 €
46,237 €
37,158 €
37,755 €
39,164 €
39,130 €
39,597 €
40,517 €
41,317 €
42,260 €
43,435 €
43,988 €
44,345 €
45,336 €
46,526 €
46,757 €
47,865 €
47,697 €
47,754 €
47,331 €
47,316 €
47,651 €
48,456 €
48,739 €
49,894 €
49,982 €
49,885 €
50,102 €
51,616 €
51,674 €
52,454 €
3,054 €
3,069 €
3,175 €
3,125 €
3,163 €
3,281 €
3,342 €
3,420 €
3,595 €
3,600 €
3,683 €
3,743 €
3,779 €
3,732 €
3,909 €
3,838 €
3,854 €
3,790 €
3,737 €
3,797 €
3,949 €
3,977 €
4,131 €
4,045 €
4,012 €
4,035 €
4,146 €
4,116 €
4,174 €
P95
43,445 €
44,037 €
45,669 €
45,579 €
46,128 €
47,276 €
48,160 €
49,279 €
50,847 €
51,364 €
51,875 €
52,980 €
54,255 €
54,359 €
55,847 €
55,523 €
55,550 €
55,021 €
54,892 €
55,347 €
56,492 €
56,822 €
58,248 €
58,175 €
58,056 €
58,333 €
60,101 €
60,052 €
61,014 €
F95-99
Mean
56,553 €
56,786 €
58,718 €
58,433 €
59,227 €
60,395 €
61,777 €
62,977 €
65,095 €
65,847 €
66,712 €
68,136 €
69,348 €
69,232 €
70,808 €
70,141 €
70,256 €
69,439 €
69,366 €
70,067 €
71,803 €
72,463 €
74,465 €
74,045 €
73,710 €
74,125 €
76,609 €
76,729 €
78,060 €
F95-99 Std
10,538 €
10,269 €
10,501 €
10,325 €
10,483 €
10,438 €
10,805 €
10,841 €
11,254 €
11,509 €
11,716 €
12,068 €
11,944 €
11,620 €
11,606 €
11,376 €
11,422 €
11,240 €
11,215 €
11,485 €
12,114 €
12,427 €
12,899 €
12,611 €
12,495 €
12,631 €
13,245 €
13,441 €
13,723 €
P99
83,079 €
82,927 €
85,336 €
84,266 €
85,500 €
86,300 €
88,954 €
89,753 €
92,893 €
94,598 €
95,677 €
98,334 €
99,056 €
98,043 €
99,649 €
98,318 €
98,942 €
97,329 €
97,350 €
98,606 €
102,520 €
103,843 €
107,180 €
106,329 €
105,592 €
106,543 €
110,894 €
111,467 €
113,781 €
F99-99.9
Mean
110,412 €
107,797 €
110,183 €
109,577 €
110,873 €
111,769 €
115,585 €
115,878 €
121,337 €
123,501 €
125,079 €
128,496 €
129,497 €
128,048 €
129,206 €
127,875 €
127,945 €
125,219 €
126,270 €
129,096 €
136,022 €
138,010 €
143,404 €
142,049 €
141,749 €
143,537 €
149,955 €
152,005 €
155,959 €
F99-99.9
Std
24,100 €
20,557 €
20,978 €
21,338 €
21,897 €
22,072 €
23,058 €
22,529 €
24,704 €
24,727 €
25,864 €
25,686 €
26,196 €
25,683 €
25,427 €
25,749 €
24,719 €
24,183 €
25,282 €
26,968 €
29,918 €
30,476 €
33,091 €
32,540 €
32,896 €
34,057 €
36,497 €
37,980 €
39,325 €
P99.9
183,050 €
167,550 €
170,093 €
172,087 €
174,628 €
177,507 €
183,025 €
180,719 €
195,690 €
195,460 €
201,399 €
201,628 €
206,521 €
201,512 €
202,710 €
205,394 €
202,017 €
197,712 €
200,441 €
210,077 €
229,015 €
234,457 €
252,082 €
247,598 €
245,381 €
253,278 €
266,729 €
273,951 €
285,527 €
F99.999.99
Mean
230,117 €
208,334 €
204,923 €
212,782 €
211,916 €
223,168 €
233,406 €
227,480 €
249,900 €
240,099 €
251,313 €
252,274 €
258,354 €
254,962 €
259,364 €
264,314 €
259,720 €
257,542 €
264,106 €
282,243 €
316,474 €
320,895 €
351,439 €
350,434 €
336,867 €
349,348 €
378,316 €
389,019 €
420,043 €
F99.999.99 Std
P99.99
41,955 €
36,030 €
30,097 €
33,001 €
30,733 €
42,255 €
45,153 €
40,056 €
51,582 €
39,109 €
44,367 €
47,804 €
48,700 €
49,264 €
54,160 €
56,028 €
47,794 €
54,421 €
56,490 €
68,806 €
80,274 €
82,686 €
92,246 €
96,936 €
87,124 €
91,678 €
106,398 €
108,801 €
128,585 €
354,629 €
329,028 €
300,252 €
306,318 €
314,561 €
355,700 €
368,471 €
348,280 €
404,474 €
364,627 €
396,400 €
414,841 €
423,901 €
412,031 €
433,983 €
445,012 €
407,830 €
433,172 €
437,931 €
505,238 €
562,880 €
589,959 €
655,473 €
658,239 €
612,332 €
653,136 €
722,456 €
748,510 €
835,934 €
F99.99-100
Mean
617,527 €
454,875 €
410,739 €
411,695 €
413,035 €
711,491 €
753,563 €
531,495 €
745,958 €
514,812 €
598,328 €
544,504 €
582,962 €
581,312 €
666,159 €
665,313 €
600,817 €
669,643 €
675,288 €
758,573 €
821,373 €
964,457 €
1,018,925 €
1,085,736 €
1,141,530 €
1,256,144 €
1,222,500 €
1,752,971 €
1,810,961 €
F99.99-100
Std
491,825 €
206,709 €
167,556 €
161,894 €
131,824 €
484,745 €
581,191 €
239,727 €
504,037 €
208,007 €
232,914 €
108,896 €
188,851 €
181,237 €
275,301 €
314,843 €
223,798 €
338,505 €
254,620 €
288,755 €
290,008 €
507,948 €
599,667 €
639,644 €
864,798 €
980,544 €
723,692 €
3,273,959 €
2,287,504 €
All Mean
19,827 €
20,279 €
21,099 €
21,246 €
21,524 €
21,980 €
22,393 €
22,708 €
23,233 €
23,393 €
23,469 €
23,854 €
24,484 €
24,666 €
24,946 €
24,771 €
24,931 €
24,841 €
24,875 €
25,101 €
25,109 €
25,223 €
25,773 €
26,074 €
26,093 €
26,287 €
27,108 €
27,311 €
27,790 €
All Std
17,485 €
15,915 €
16,009 €
15,989 €
16,125 €
18,010 €
18,849 €
17,397 €
19,471 €
18,248 €
18,901 €
19,020 €
19,450 €
19,304 €
20,016 €
20,064 €
19,625 €
19,728 €
19,758 €
20,541 €
21,821 €
23,009 €
24,310 €
24,529 €
25,206 €
26,395 €
26,324 €
43,337 €
37,294 €
N
535 292
545 229
534 515
549 794
542 101
521 238
485 279
479 365
483 321
486 983
484 805
508 852
527 896
534 834
518 688
513 955
522 383
527 345
533 517
551 889
533 422
560 247
580 405
1 182 443
1 186 862
1 174 623
1 178 154
1 245 212
1 269 372
Table S2. Thresholds, means and standard estimates of yearly gross wages for different fractiles of the distribution (Exhaustive files). Constant
2007 euros.
Year
1994
1995
1996
1997
1998
1999
2000
2001
2001
2002
2003
2004
2005
2006
2007
Selection
threshold
6,670 €
6,757 €
6,864 €
6,979 €
6,931 €
7,218 €
7,256 €
7,395 €
7,395 €
7,495 €
7,627 €
7,883 €
7,747 €
7,525 €
7,604 €
F00-90
Mean
20,913 €
21,135 €
21,183 €
21,189 €
21,043 €
21,427 €
21,191 €
21,526 €
21,674 €
21,956 €
21,640 €
21,723 €
21,772 €
21,975 €
22,396 €
F00-90
Std
8,200 €
8,186 €
8,148 €
8,172 €
8,116 €
8,369 €
8,388 €
8,429 €
8,841 €
8,900 €
9,083 €
8,774 €
8,813 €
8,967 €
9,104 €
P90
42,431 €
42,642 €
42,581 €
42,611 €
42,318 €
43,430 €
43,284 €
44,039 €
45,521 €
45,939 €
46,378 €
45,291 €
45,481 €
45,896 €
46,708 €
F90-95
Mean
48,062 €
48,251 €
48,111 €
48,108 €
47,777 €
49,174 €
49,046 €
49,981 €
51,794 €
52,175 €
52,503 €
51,363 €
51,598 €
52,019 €
52,935 €
F90-95
Std
3,764 €
3,766 €
3,718 €
3,717 €
3,692 €
3,849 €
3,850 €
3,969 €
4,190 €
4,152 €
4,044 €
4,052 €
4,087 €
4,099 €
4,179 €
P95
55,753 €
55,850 €
55,617 €
55,590 €
55,208 €
56,978 €
56,833 €
58,014 €
60,291 €
60,590 €
60,640 €
59,584 €
59,905 €
60,358 €
61,452 €
F95-99
Mean
69,860 €
70,073 €
69,668 €
69,624 €
69,145 €
71,621 €
71,715 €
73,493 €
76,686 €
76,823 €
76,147 €
75,486 €
76,126 €
76,777 €
78,312 €
F95-99 Std
11,028 €
10,990 €
10,893 €
10,896 €
10,821 €
11,502 €
11,702 €
12,275 €
13,020 €
12,906 €
12,349 €
12,697 €
12,992 €
13,181 €
13,585 €
P99
97,262 €
97,298 €
96,692 €
96,708 €
96,043 €
100,559 €
101,263 €
104,654 €
109,881 €
109,805 €
107,805 €
108,083 €
109,528 €
110,839 €
113,613 €
F99-99.9
Mean
124,806 €
124,099 €
123,473 €
124,334 €
123,479 €
131,037 €
133,220 €
139,155 €
146,385 €
146,280 €
143,795 €
144,996 €
147,842 €
150,363 €
155,574 €
F99-99.9
Std
23,670 €
23,116 €
23,204 €
24,126 €
23,960 €
27,210 €
29,001 €
31,411 €
33,065 €
33,016 €
32,749 €
33,361 €
34,910 €
36,323 €
39,058 €
P99.9
195,372 €
192,742 €
192,977 €
196,959 €
195,604 €
216,320 €
225,847 €
240,293 €
252,935 €
251,250 €
247,394 €
250,379 €
259,262 €
266,399 €
282,787 €
F99.999.99
Mean
247,703 €
245,887 €
247,821 €
256,797 €
255,030 €
296,305 €
314,418 €
341,918 €
356,328 €
350,605 €
340,690 €
350,292 €
366,355 €
378,753 €
414,348 €
F99.999.99 Std
47,337 €
47,083 €
48,920 €
53,920 €
53,549 €
74,997 €
83,948 €
96,814 €
98,404 €
94,937 €
86,488 €
94,943 €
102,018 €
108,988 €
131,333 €
P99.99
392,474 €
389,209 €
396,729 €
424,775 €
421,853 €
538,701 €
597,029 €
653,796 €
673,132 €
661,717 €
619,859 €
660,466 €
701,003 €
739,876 €
866,596 €
F99.99-100
Mean
594,876 €
570,918 €
580,261 €
640,994 €
636,584 €
860,305 €
977,859 €
985,567 €
1,057,729 €
1,158,246 €
1,143,812 €
1,232,770 €
1,281,250 €
1,421,522 €
1,682,324 €
F99.99-100 Std
493,287 €
280,595 €
267,212 €
289,023 €
287,034 €
426,292 €
506,115 €
411,276 €
604,312 €
870,210 €
975,510 €
1,113,680 €
1,336,719 €
1,647,971 €
1,631,479 €
All Mean
25,425 €
25,632 €
25,650 €
25,674 €
25,498 €
26,140 €
25,973 €
26,471 €
26,908 €
27,190 €
26,863 €
26,882 €
27,009 €
27,286 €
27,878 €
All Std
19,651 €
19,078 €
19,009 €
19,386 €
19,253 €
21,507 €
22,562 €
23,216 €
24,905 €
26,013 €
26,148 €
27,062 €
28,642 €
31,147 €
33,067 €
N
11 439 684
11 253 327
11 233 725
11 429 251
11 429 251
11 942 916
12 400 411
12 670 098
15 146 231
15 160 086
16 066 991
15 950 337
16 321 586
16 724 583
16 921 903
29